Jointly optimal dereverberation and beamforming
This is an incremental analysis for audio signal processing researchers, clarifying the source of performance gains in dereverberation and beamforming methods.
The paper tackled the problem of understanding why a previously proposed convolutional beamformer outperforms a cascade approach for simultaneous denoising and dereverberation, and found that its superiority stems from a specific beamforming component called wMPDR.
We previously proposed an optimal (in the maximum likelihood sense) convolutional beamformer that can perform simultaneous denoising and dereverberation, and showed its superiority over the widely used cascade of a WPE dereverberation filter and a conventional MPDR beamformer. However, it has not been fully investigated which components in the convolutional beamformer yield such superiority. To this end, this paper presents a new derivation of the convolutional beamformer that allows us to factorize it into a WPE dereverberation filter, and a special type of a (non-convolutional) beamformer, referred to as a wMPDR beamformer, without loss of optimality. With experiments, we show that the superiority of the convolutional beamformer in fact comes from its wMPDR part.